Re: st: Identifying subfamilies (2nd+ marriages) within records of families

[Nick Cox <njcoxstata@gmail.com>]

Blonde moment!!!

Sounds like

bysort family (reg) : gen familynum = sum(reg != reg[_n-1]) - 1

With this definition, I am assuming that -family- is string (else how
can a value be empty?) and  the wife herself is labelled 0.

There would be a small tweak  needed if -family- were really numeric.

Nick

On Wed, Feb 1, 2012 at 10:53 PM, Anna Reimondos <areimondos@gmail.com> wrote:

> I have a dataset of families which looks something like this:
>
> The first variable is a family id, then there is a registration
> number, and a variable identifying whether the person is the wife,
> husband or child.
> In the example below there are two families. In the first family there
> is a wife, husband and kids. In the second family the woman has had 2
> marriages. In the first she had 2 children, and in the second marriage
> she had one child. What I would like to do is create a variable called
> 'familynum' which will number the different families within the main
> family.
>
> NB: There is no registration or familynum for the woman.
>
>     family   reg        person        familynum
>        1                       Wife    (1)
>        1       693             Husband(2)      1
>        1       693             Child   (3)             1
>        1       693             Child   (3)             1
>        2                       Wife    (1)
>        2       542             Husband(2)      1
>        2       542             Child   (3)             1
>        2       542             Child   (3)             1
>        2       879             Husband(2)      2
>        2       879             Child   (3)             2
>
>
> I have done similar things before, and I know the solution is probably
> very obvious...but I am having a blonde moment.

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/